FWIW, I agree with Martin on the points he defends below. And amsmath is bound to be incompatible with some latex packages. This is latex, after all. OTOH, not wanting ams* is not a matter of availability, since the ams classes are in reauired/ on CTAN, and I think we can consider that they are part of any standard latex installation.
JMarc Martin Vermeer wrote: > 1) It *is* reliable for the things "dummies" do -- except perhaps for > bugs that just need to get fixed then. Non-dummies are on their own > anyway. > > 2) The marginal code added for AMS as against only validating other > things is small and uncontroversial. And will become smaller still, > the more optional packages we are going to support (wasn't that the > idea?) > > 3) No. My solution (auto-validate with manual negative override) needs > one button, just like yours (auto-include with manual negative override > IIUC). The only real difference is that it *does* auto-validate before > even looking at the button.
